Versatile Naphthalimide Tetrazines for Fluorogenic Bioorthogonal Labelling† Cite This: DOI: 10.1039/D1cb00128k Ab Ab Cd E Marcus E

View Article Online PAPER View Journal

Versatile naphthalimide tetrazines for fluorogenic bioorthogonal labelling† Cite this: DOI: 10.1039/d1cb00128k ab ab cd e Marcus E. Graziotto, Liam D. Adair, Amandeep Kaur, Pauline Ve´rite´, Sarah R. Ball,c Margaret Sunde, cd Denis Jacquemin e and Elizabeth J. New *abd

Fluorescent probes for biological imaging have revealed much about the functions of biomolecules in health and disease. Fluorogenic probes, which are fluorescent only upon a bioorthogonal reaction with a specific partner, are particularly advantageous as they ensure that fluorescent signals observed in biological imaging arise solely from the intended target. In this work, we report the first series of naphthalimide tetrazines for bioorthogonal fluorogenic labelling. We establish that all of these compounds can be used for imaging through photophysical, analytical and biological studies. The best candidate was Np6mTz, where the tetrazine ring is appended to the naphthalimide at its 6-position via a

Creative Commons Attribution-NonCommercial 3.0 Unported Licence. phenyl linker in a meta configuration. Taking our synthetic scaﬀold, we generated two targeted variants, LysoNpTz and MitoNpTz, which successfully localized within the lysosomes and mitochondria Received 10th June 2021, respectively, without the requirement of genetic modification. In addition, the naphthalimide tetrazine Accepted 24th June 2021 system was used for the no-wash imaging of insulin amyloid fibrils in vitro, providing a new method that DOI: 10.1039/d1cb00128k can monitor their growth kinetics and morphology. Since our synthetic approach is simple and modular, these new naphthalimide tetrazines provide a novel scaﬀold for a range of bioorthogonal tetrazine- rsc.li/rsc-chembio based imaging agents for selective staining and sensing of biomolecules.

This article is licensed under a Introduction have attracted much attention due to their ability to quench fluorescence via both through-space Fo¨rster resonance energy Fluorescent probes are a mainstay of molecular imaging, transfer (FRET)9 and through-bond energy transfer (TBET) 10–12 Open Access Article. Published on 05 July 2021. Downloaded 9/23/2021 12:29:10 PM. providing previously inaccessible information about the processes. They have been shown to react rapidly complex chemistry of biomolecules, cells and organisms.1–3 (second-order rate constants up to 106 MÀ1 sÀ1)4 in inverse Recent advances in bioorthogonal chemistry have allowed for electron demand Diels Alder (IEDDA) reactions with a range of the development of fluorogenic probes, for which fluorescence strained dienophiles such as trans-cyclooctenes8 and intensities dramatically increase upon a click reaction with a bicyclononynes.13 The IEDDA reaction of a tetrazine with a 4,5 bioorthogonal partner. These fluorogenic probes have been strained cycloalkyne produces a pyridazine with only N2 as a by- extensively used to image biological structures, revealing their product (Fig. 1A). Fluorescence is restored to the fluorophore as significance during health and disease.6 pyridazines do not quench fluorescence through energy trans- Of the suite of bioorthogonal reactions developed for studies fer and hence the tetrazine-BCN ligation is an excellent strategy of biological processes, the tetrazine ligation has been exten- for fluorogenic labelling.14 sively utilized for fluorogenic probes.7,8 The 1,2,4,5-tetrazines Many tetrazine-containing fluorogenic probes have been synthesized with emission wavelengths spanning the visible and infrared spectrum, commonly employing coumarin,15 a The University of Sydney, School of Chemistry, NSW, 2006, Australia. fluorescein,16 rhodamine,17 cyanine,18 BODIPY12 and other E-mail: [email protected] 19–25 b Australian Research Council Centre of Excellence for Innovations in Peptide and commercial and novel scaﬀolds. All of these have been Protein Science, The University of Sydney, NSW, 2006, Australia utilized in confocal microscopy, and some for super resolution 26–28 c The University of Sydney, School of Medical Sciences, Faculty of Medicine and imaging. Typically, these reports require the genetic mod- Health, NSW, 2006, Australia ification of a native protein to incorporate a bioorthogonal d The University of Sydney Nano Institute (Sydney Nano), The University of Sydney, reactive group and this method only provides information on NSW, 2006, Australia e CEISAM Lab, CNRS, Universite´ de Nantes, Nantes, France the localization of that macromolecule. The few notable excep- † Electronic supplementary information (ESI) available. See DOI: 10.1039/ tions, where fluorogenic tetrazines have been used for targeted 2+ d1cb00128k or analyte sensing, include: a Mg fluorescent sensor with a

In this work, we describe the efficient synthesis and photophysical properties of the first series of naphthalimide tetrazines that can be used for biological imaging. Using the optimized scaffold from these studies, we developed live cell fluorogenic organelle-targeted probes that do not require genetic modification or antibody-based stains. In addition, the naphthalimide tetrazines were used to label insulin amyloid fibrils in vitro, without washing and without affecting their growth kinetics.

Results and discussion Design and synthesis In our previous work with substituted naphthalimides, we identified that derivatives with substituents installed at the 3- and 6-position on the naphthalene core exhibited the best photophysical properties for imaging.32 We chose to conjugate the tetrazine at the 3- and 6-positions with a phenyl ring to ensure TBET quenching.12 FRET quenching is also expected to occur in this molecule as the naphthalimide and tetrazine are in close proximity. As FRET is directional, it was anticipated that the configuration across the phenyl ring would cause Creative Commons Attribution-NonCommercial 3.0 Unported Licence. diﬀerent degrees of quenching.15 Hence, we investigated molecules in which the tetrazine was installed in meta and para Fig. 1 (A) Fluorescent tagging of biomolecules using the tetrazine liga- configurations on the phenyl ring, relative to the naphthali- tion. Tetrazines quench appended fluorophores via FRET and TBET and mide. Four compounds were designed to explore the relative after reaction with a cycloalkyne, form a pyridazine which does not degree of quenching and subsequent fluorescence turn-ons of quench fluorescence. (B) Synthesis of the four desired naphthalimide tetrazines through the coupling of 1a or 1b with 2a or 2b. naphthalimide tetrazines. A convergent synthetic route was envisaged, employing a convergent cross-coupling between bromo-naphthalimides and

This article is licensed under a tetrazine for organelle-localized Mg2+ detection;29 a fluorogenic tetrazine-aryl-boronate esters as the final step. 3-Bromo- reaction to quantify endocytosis of antibody conjugates;30 and using naphthalimide (1a) and 6-bromo-naphthalimide (1b) were tetrazines as a phototrigger to activate organelle-targeted stains.31 obtained from anhydride intermediates in moderate yields using previously reported conditions (Scheme S1, ESI†).42 The Open Access Article. Published on 05 July 2021. Downloaded 9/23/2021 12:29:10 PM. We ascribe the lack of development in this area due to the challenge of finding fluorescent moieties which have synthetic handles that tetrazine-aryl-boronate esters were synthesized using condi- can be readily decorated with sensing or targeting groups. tions recently reported by Mao et al. for the thiol-catalyzed 43 The 4-amino-1,8-naphthalimides are a class of fluorophores formation of tetrazines from aryl nitriles. Hence, bromophe- for which tetrazine conjugates for bioimaging applications have nyltetrazines were then prepared from 3-bromobenzonitrile not been reported to date. These fluorophores have great and 4-bromobenzonitrile respectively. Miyaura borylation of potential for bioimaging due to their brightness, large Stokes these bromide intermediates aﬀorded the boronate esters 2a shifts and good photostability.32 In addition, they can be and 2b in good yields (Scheme S1, ESI†). The two boronate readily synthetically modified at the imide, 4-amino-position esters were then coupled to the two bromo-naphthalimides in and 3-, 5- or 6-positions of the naphthalene core.32,33 There are all combinations using standard Suzuki cross-coupling condi- some reports of fluorogenic naphthalimides for biological tions, aﬀording the four desired naphthalimide tetrazines in imaging and protein labelling, with the fluorogenic changes moderate yields (Fig. 1B). arising from click reactions involving azides and alkynes,34 SNAP tags,35 sydnones,36 and oximes.37 However, there are no Photophysical properties reports of fluorogenic 4-amino-naphthalimides incorporating With the four candidates in hand, we first established that tetrazines for biological imaging. To the best of our knowledge, none of the unreacted tetrazine products was significantly the only two naphthalimide tetrazines reported to date were fluorescent. While all compounds showed significantly developed for electrochemical applications and not applied to quenched fluorescence, a weak emission band was observed biological imaging. They were unsuitable for biological applica- around 530 nm in ethanol for all compounds (Table 1). To tion as these reports used a 1,8-naphthalimide that has signifi- quantify the weak fluorescence, we attempted to measure the cantly shorter excitation and emission wavelengths compared absolute quantum yields of the tetrazines in ethanol. All dyes, to 4-amino-substituted naphthalimides.38–41 with the exception of Np3pTz, had a fluorescence quantum

Table 1 Fluorescent properties of naphthalimide tetrazines and the corresponding pyridazine reaction products in absolute ethanol

lex lem e B À1 À1 À1 À1 (nm) (nm) (M cm ) ff (M cm ) Np3mTz 445 528 9500 o0.01 a Np3mPz 449 539 8800 0.24 2100 Np3pTz 451 533 9700 0.027 260 Np3pPz 449 545 11 000 0.25 2700 Np6mTz 460 530 6700 o0.01 a Np6mPz 461 540 8900 0.34 3000 Np6pTz 463 533 11 000 o0.01 a Np6pPz 460 537 11 000 0.25 2500

a Brightness (e Â ff) could not be accurately calculated.

yield below the detection limit of the instrument. The estimated brightness for each tetrazine was less than 260 MÀ1 cmÀ1,in contrast to dyes used for molecular imaging that typically exhibit much greater brightness in the range 103–106 MÀ1 cmÀ1.1 Fluores- cence emission was not observed in HEPES buffer (Table S1, ESI†), which indicates that these probes would not fluoresce in aqueous cellular media and are therefore suitable for no-wash imaging protocols. Satisfied that all of our tetrazine candidates exhibited sufficiently quenched fluorescence, particularly in aqu-

Creative Commons Attribution-NonCommercial 3.0 Unported Licence. eous media, we then sought to determine whether each naphthalimide tetrazine produced a fluorogenic response after click Fig. 2 (A) Schematic of fluorogenic reaction between naphthalimide reaction. (Naph) tetrazines after reaction with BCN to form pyridazine products in situ. (B) Photograph of vials of Np6mTz (left) and Np6mPz (right) in For initial studies, the commercially available reagent dichloromethane, under 365 nm light. (C) Emission spectra of 5 mM (1R,8S,9s)-bicyclo[6.1.0]non-4-yn-9-ylmethanol (BCN) was used solutions of naphthalimide tetrazines (orange) and the corresponding as the click partner. The relevant tetrazine was incubated with pyridazine products (black). Numbers indicate the fold turn-on as deter- five equivalents of BCN in DMSO for 10 min (Fig. 2A). After mined by integration of the spectra. reaction with BCN, a significant change in the color of the solution was observed for all naphthalimide tetrazines (Fig. 2B), observed for all compounds, with fluorescence emission red- This article is licensed under a with corresponding shifts in the absorption spectra and disap- shifting with the polarity of the solvent (Fig. S11, ESI†). All pearance of shoulder absorbance bands (500–550 nm) that corre- pyridazine products exhibited a brightness greater than spond to the n–p* transition of tetrazines (Fig. S1, ESI†). This was 103 MÀ1 cmÀ1, which is considered to be suﬃciently bright

Open Access Article. Published on 05 July 2021. Downloaded 9/23/2021 12:29:10 PM. consistent with the formation of the naphthalimide pyridazine for cellular imaging,1 with Np6mPz being the brightest deriva- products, which we named NpxxPz (Scheme S2, ESI†). We verified tive. Since the cellular environment is not homogenous in the formation of the pyridazines using liquid-chromatography polarity, we evaluated the fold turn-on of the naphthalimide mass spectrometry (LCMS) (Fig. S2–S5, ESI†) and the masses of tetrazines after reaction with BCN using integrated fluores- these products were confirmed with high resolution mass spectro- cence emission intensities in the same solvents employed for metry (HRMS) (Fig. S6–S9 and Table S2, ESI†). These studies solvatochromism studies (Fig. S12, ESI†). Np6mTz exhibited a indicated that the bioorthogonal reaction had proceeded to 200-fold turn-on upon reaction with BCN in ethanol and had completion. The second-order rate constants of the click reaction the highest fold enhancements in all solvents. It was interesting were determined experimentally using the fluorescence increase to note that despite the significant diﬀerences between the observed after treatment with BCN (Fig. S10, ESI†). Rate constants compounds across the range of solvents, all of the compounds À1 À1 were in the range of 4–8 M s in a 1:1 acetonitrile:water had a similar 70–130-fold turn-on in HEPES buffer. We ascribe mixture, which is comparable to other reported tetrazine-BCN this observation to the extremely low fluorescence of all the 44 4 reactions and sufficiently fast for live-cell labelling. naphthalimide tetrazines in aqueous solvents, reinforcing their We compared the emission spectra of equal concentrations suitability for no-wash cellular imaging. The combined photo- of the tetrazines and corresponding pyridazine products, with physical data indicated that Np6mTz is the best precursor for strong fluorescence increases observed after the click reaction tetrazine-based fluorogenic probes for biological applications. with BCN (Fig. 2C). Solutions of the pyridazine products in ethanol showed strong fluorescence, with quantum yields between 0.24 and 0.34. The 6-position derivatives exhibited First-principle calculations slightly red-shifted excitation maxima compared to the To confirm our experimental observations of the photophysical 3-position, but with similar emission maxima (Table 1). As properties of the naphthalimide tetrazines, we have performed expected for ICT fluorophores, a strong solvatochromism was ab initio calculations relying on a mixed protocol combining

time-dependent density functional theory (TD-DFT) and of the tetrazine and naphthalimide that are significantly dif- second-order Coupled Cluster approaches (see the ESI† for ferent in the two dyes. With this data, we rationalized key details). A comparison between the theoretical and experi- results from our photophysical experiments and were confident mental values for excitation energies reveals errors in the usual in the choice of Np6mTz as most suitable for further range for such a model (Fig. S13 and Table S3, ESI†).45 For all applications. naphthalimide tetrazines, TD-DFT theory indicates that the first transition is localized on the tetrazine with a vertical Investigating the fluorogenic reaction in cells transition energy of approximately 2.30 eV and a trifling oscil- We next sought to confirm that the fluorogenic reaction p lator strength, consistent with a dark n– * excitation, whereas between the naphthalimide tetrazines and BCN could be the second transition at approximately 3.01 eV corresponds to a observed in living cells. Before carrying out imaging studies, naphthalimide-centered excitation, which is very bright A549 cells were exposed to high concentrations of the tetrazines 4 (f 0.3). As can be seen in Fig. 3 for Np3pTz, the two (25 mM), BCN (250 mM) or a combination of the two for 3 h to excitations are not overlapping at all, and the second excitation determine whether these compounds were cytotoxic. Pleas- presents a strong charge transfer character, with the amino ingly, we observed no significant changes in cell viability during group acting as the main donor (large blue lobe in Fig. 3 this time (Fig. S14, ESI†). The intracellular fluorescence of cells center). While the S0–S1 transition induces no change in the treated with Np3mTz as a model tetrazine for 20 min or 90 min, total dipole moment, the S0–S2 excitation of Np3pTz is accom- with and without washing, was similar for all conditions, panied by an increase of the dipole by +5.36 D, consistent with confirming that the dyes could be used for no-wash labelling the measured positive solvatochromism. When the Np3pPz and imaging (Fig. S15, ESI†). We then compared the fluores- structure is formed, the lowest transition disappears and the cence of cells treated with the tetrazines and the pyridazine S0–S1 transition becomes equivalent to the second transition in products. Cells treated with the tetrazines (1 mM, 20 min) the tetrazine derivative. Qualitatively equivalent results are exhibited minimal fluorescence compared to cells treated with

Creative Commons Attribution-NonCommercial 3.0 Unported Licence. obtained for the four derivatives. We have also determined the corresponding pyridazines (Fig. S16, ESI†). Importantly, the the S1 and S2 minimal structures for all four naphthalimide pyridazines gave measurable signals at low concentrations derivatives, allowing direct calculation of the fluorescence (1 mM) and low laser powers (o1.0%). The only tetrazine that energies. In addition, the state ordering is not affected by the gave any observable fluorescence in cells was Np3pTz, consis- † geometrical relaxation (Table S4, ESI ). Therefore, the weak tent with its higher quantum yield. emission from the tetrazine derivatives is due to a residual We then confirmed that the reaction partners underwent the radiative deexcitation from the S2 state, with the very strong fluorogenic click reaction inside live cells. Strong fluorescence quenching being explained by FRET from this S2 state to the was observed in cells that were first treated with a naphthali- lower dark S1 state. Finally, to understand why Np3pTz is mide tetrazine (1 mM, 30 min), followed by BCN (50 mM, 1 h) This article is licensed under a significantly less quenched than Np3mTz, we have determined (Fig. 4), confirming that the bioorthogonal reaction occurs the excitation energy transfer (EET) coupling constant between within cells and that the fluorescent product is suﬃciently the S1 and S2 states at the S2 minimal geometry. As EET is bright for confocal microscopy. We next evaluated how fast Open Access Article. Published on 05 July 2021. Downloaded 9/23/2021 12:29:10 PM. significantly dependent on the relative orientations of the two this reaction could occur within cells, and based on our chromophores, we considered several conformations and aver- photophysical data, chose Np6mTz as the candidate for testing. aged our results. The TD-DFT calculations found a significantly smaller coupling value in Np3pTz (1.63 meV) than in Np3mTz (2.49 meV), consistent with a stronger quenching of the emission in the latter. As the relative energies of the various states are almost unaffected by the various linkages, this effect is a result of the orientation of the two transition dipole moments

Fig. 3 Density diﬀerence plots for selected excited states of Np3pTz and Fig. 4 Representative images of A549 cells stained with Np3mTz, Np3pPz. The blue and red regions represent regions of increase and Np3pTz, Np6mTz or Np6pTz (1 mM, 30 min) followed by incubation with decrease of electron density upon photon absorption, respectively. Text BCN (50 mM, 1 h). lex = 488 nm, lem = 510–610 nm. Scale bars represent indicates the computed transition energies and oscillator strengths (f). 20 mm.

Cells were first incubated with Np6mTz, washed, and then ff = 0.27 in EtOH) had similar absorbance and fluore- incubated with imaging media or BCN-doped imaging media scence profiles to the 6-position naphthalimide pyridazines and imaged every 2 min for 30 min (Fig. S17, ESI†). Bright (Fig. S19–S21, ESI†). Furthermore, high doses of LysoNpTz fluorescence was observed within 20 min, indicating that the and MitoNpTz were non-toxic to A549 cells (Fig. S22, ESI†). BCN partner rapidly enters the cell and reacts with the The probes also showed a fluorogenic response in cells within tetrazine. 30 min upon co-treatment with BCN (50 mM) (Fig. S23, ESI†). With confidence that the targeted naphthalimide click reac- Targeted naphthalimide tetrazines tions had similar photophysical and biological properties to the Having optimized the general scaﬀold for fluorogenic imaging, parent compounds, we next evaluated the ability of the probes we then sought to utilize the synthetic versatility of the system to label their respective organelles. To verify their diﬀerences in for targeting the tetrazines to cellular organelles. We envisaged localization, both naphthalimide tetrazines were incubated that targeting the naphthalimide tetrazines to the lysosomes with BCN and LysoTracker Deep Red (LTDR) or MitoTracker and mitochondria with appropriate targeting groups would Deep Red (MTDR). Fluorescence was observed in punctate produce fluorescent responses that could be correlated with regions around the nucleus in cells treated with LysoNpTz established cellular trackers. We chose a morpholine targeting and BCN, which corresponded to the lysosomes (Fig. 5B) group for the lysosomes, which accumulates based on the and did not significantly colocalize with the mitochondria protonation of the tertiary amine inside the acidic environment (Fig. S24, ESI†). In contrast, cells treated with MitoNpTz and of the lysosomes; and a triphenylphosphonium group for BCN exhibited web-like fluorescent structures that correlated to targeting the mitochondria, which bears a positive charge that the mitochondria (Fig. 5C) and did not exhibit fluorescence in accumulates due to the negative membrane potential of the the lysosomes (Fig. S24, ESI†). Taken together, it is clear that mitochondria. incorporating these targeting groups increases the selectivity of Both targeted variants were synthesized from 3-bromo-5- the probes for the organelles of live cells, without the need for

Creative Commons Attribution-NonCommercial 3.0 Unported Licence. nitro-1,8-naphthalic anhydride; the intermediate previously genetic modification of a specific protein. used for synthesis of 1b, the precursor of Np6mTz. The lysosome-targeted naphthalimide, LysoNpTz, was synthesized Monitoring insulin fibrillation with fluorogenic over three steps in 46% overall yield (Scheme S3, ESI†). The naphthalimides mitochondria-targeted variant, MitoNpTz, was synthesized in Having demonstrated that we can readily modify the naphtha- five steps in 11% overall yield, with the additional steps limide tetrazines for organelle labelling, we sought to use the permitting the introduction of the charged targeting group in system as a means to investigate other biological structures of the final step (Scheme S4, ESI†). Both of these naphthalimide interest. Amyloid fibrils are implicated in a range of neurode- tetrazines were reacted with BCN in situ as before, generating generative disorders, and it is therefore essential to be able to This article is licensed under a the pyridazine products LysoNpPz and MitoNpPz (Fig. 5A) image their formation and localization.46 The visualization of whose identities were confirmed by LRMS (Fig. S18, ESI†). these structures is challenging due to the limitations in speci- À1 À1 47 LysoNpPz (lem = 532 nm, e = 12 000 M cm , ff = 0.30 in ficity and labelling eﬃciency of dyes like thioflavin T (ThT),

Open Access Article. Published on 05 July 2021. Downloaded 9/23/2021 12:29:10 PM. À1 À1 EtOH) and MitoNpPz (lem = 543 nm, e = 4900 M cm , and as such, genetically-modified amyloid-forming proteins

Fig. 5 Targeted naphthalimide tetrazines. (A) Reaction of LysoNpTz with BCN to form LysoNpPz and reaction of MitoNpTz with BCN to form MitoNpPz. Representative images of A549 cells dosed with (B) LysoNpTz (2.5 mM) and BCN (2.5 mM) and LysoTracker Deep Red (LTDR) (50 nM) or

(C) MitoNpTz (10 mM) and BCN (10 mM) and MitoTracker Deep Red (MTDR) (100 nM). Green channel (naphthalimides): lex = 488 nm, lem = 510–610 nm.

Red channel (trackers): lex = 640 nm, lem = 650–750 nm. Scale bars represent 20 mm.

tagged with fluorescent proteins are often used in experiments. imaging of insulin fibrils, so the monomers were used as a However, the genetic modification of amyloid-forming proteins mixture in all subsequent experiments. presents multiple challenges associated with cloning and Amyloid fibrils were prepared from insulin and insulin-BCN altered physiological and aggregation profile of the amyloid- using established procedures and imaged using total internal forming proteins.48,49 An alternative approach involving label- reflection microscopy (TIRF). The untreated insulin-BCN fibrils ing with small-molecule fluorophores in vitro may disturb fibril were non-fluorescent, while unlabeled insulin fibrils incubated growth, leading to models that do not accurately reflect with Np6mTz showed very weak fluorescence (Fig. 6A and B), their physiology.50 We therefore investigated whether the potentially due to accumulation of Np6mTz in the hydrophobic BCN-tetrazine labelling reaction could overcome these chal- environment of the fibrils. The low fluorescence confirms that lenges as the BCN modification is smaller than most fluoro- Np6mTz does not undergo a click reaction with untagged phores used for imaging. insulin. In contrast, the insulin-tagged BCN fibrils after incuba- Insulin was chosen for these studies as it is widely used as a tion with Np6mTz resulted in a strong fluorescence signal model to probe the mechanisms of amyloid fibril formation, as consistent with formation of the fluorescent pyridazine. The insulin amyloids exhibit the characteristic cross-b structure images show a web-like network morphology, as expected for found in the fibrils of disease-relevant amyloids.51 Monomeric insulin fibrils (Fig. 6C). This demonstrates that the naphthali- insulin was treated with excess BCN 4-nitrophenyl ester mide tetrazine click reaction occurs in vitro and can be applied (Scheme S5, ESI†) that was expected to react with one, two or for imaging amyloid protein assemblies. Importantly, the use of three of the primary amine sites on the insulin monomer (two the non-fluorescent Np6mTz enabled no-wash imaging experi- N-termini and a lysine residue). The tagged insulin-BCN was ments. Colocalization experiments with AmyTracker 680, a isolated using size exclusion column purification with 20 mM commercial amyloid marker, provided further confirmation glycine HCl (pH 2.0). MALDI-TOF-MS analysis was used to that the tetrazine click reaction is labeling the amyloid fibrils confirm that the mixture contained insulin tagged with one, (Fig. 6D–F).

Creative Commons Attribution-NonCommercial 3.0 Unported Licence. two or three BCN molecules (Fig. S25, ESI†). The number of To investigate whether BCN-labelling of insulin results in BCN molecules per insulin monomer does not aﬀect the altered amyloid assembly or properties, we measured the This article is licensed under a Open Access Article. Published on 05 July 2021. Downloaded 9/23/2021 12:29:10 PM.

Fig. 6 Investigating BCN-labelled insulin. Representative TIRF images of (A) insulin-BCN fibrils (0.5 mg mLÀ1) with no dye, (B) native insulin fibrils À1 À1 (2 mg mL ) incubated with Np6mTz (2 mM) and (C) insulin-BCN fibrils (0.5 mg mL ) incubated with Np6mTz (2 mM), with lex = 473 nm. Representative

TIRF images of the (D) green channel (lex = 473 nm), (E) red channel (lex = 561 nm) and (F) merge of green and red channels of insulin-BCN fibrils treated with Np6mTz (2 mM) and AmyTracker 680 (1 mM). (G) Thioflavin T fibril assembly kinetics assay. Insulin, insulin-BCN and insulin-TAMRA were incubated in glycine buﬀer (20 mM, pH 2) with thioflavin T, and fluorescence intensity monitored every 7.5 min over 500 min. Data presented as mean Æ SD of triplicate values from one experiment. (H) Representative TIRF image of insulin-BCN fibrils treated with Np6mTz (2 mM) after 500 min. (I) Representative TIRF image of insulin-TAMRA fibrils after 500 min. (J) Mean fibril area from TIRF images at the times indicated. Data presented as mean Æ SD for 25 areas of interest from one experiment. All scale bars represent 5 mm.

assembly kinetics of monomeric insulin and insulin-BCN using functionalizable fluorescent sensors that can report on biolo- a conventional ThT-based assay in which formation of amyloid gical changes at the sites of interest. fibrils is signaled by increased ThT fluorescence at 485 nm. Insulin labelled with the rhodamine derivative TAMRA (insulin- TAMRA) was also synthesized and tested to compare the effect Author contributions of labelling insulin with a larger fluorophore (Fig. S26, ESI†). We observed no significant differences between the ThT assem- The manuscript was written through contributions of all bly profiles of insulin and insulin-BCN as they formed fibrils authors. All authors have given approval to the final version (Fig. 6G), while insulin-TAMRA showed a delayed and attenu- of the manuscript. ated oligomerization. This demonstrated that labelling insulin monomers with BCN does not perturb their amyloid assembly or subsequent ThT binding, unlike insulin-TAMRA. To further Conflicts of interest probe the differential fibril assembly profiles, we imaged the There are no conflicts to declare. growth of insulin-BCN and insulin-TAMRA fibrils and measured the area covered by the fibrils from these images. Samples of insulin-BCN fibrils (treated with Np6mTz)or insulin-TAMRA fibrils were extracted at different intervals Acknowledgements (50–500 min) and imaged. TIRF imaging of the fibrils demon- The authors would like to acknowledge the Australian Research strated a significant increase in the extent of the fibrillar Council (DP180101353, DP180101897 and DP200102463) for network over time (Fig. S27 and S28, ESI†). The images clearly funding, the Westpac Scholars Trust for a Research Fellowship demonstrate formation of a less dense fibrillar network in the (EJN), the University of Sydney for a SOAR Fellowship (EJN) a case of insulin-TAMRA (Fig. 6I) compared to insulin-BCN USyd Fellowship (AK), and the John A Lamberton Scholarship

Creative Commons Attribution-NonCommercial 3.0 Unported Licence. (Fig. 6H), particularly noticeable in images collected after (MEG), and the Australian Government for Research Training 150 min. The lower density of insulin-TAMRA fibrils compared Program Scholarships (MEG, SRB). We acknowledge the scien- to insulin and insulin-BCN fibrils was also confirmed with TEM tific and technical assistance of Sydney Analytical, the Mass imaging (Fig. S29, ESI†) and by analysis of the area covered by Spectrometry Facility at the School of Chemistry and the insulin-BCN and insulin-TAMRA fibrils (Fig. 6J). We have there- Australian Microscopy and Microanalysis Research Facility at fore been able to demonstrate that the tetrazine-BCN reaction is the Australian Centre for Microscopy and Microanalysis a valuable strategy for monitoring the progression of amyloid (ACMM). PMV and DJ thank the ANR for support in the frame- fibril formation. The fact that the BCN label does not perturb work of the GeDeMi grant. This research used the computa- fibril formation, while the TAMRA label does, suggests that tional resources of CCIPL (Centre de Calcul Intensif des Pays de This article is licensed under a post-aggregation functionalization with the fluorophore is a la Loire). preferable strategy for fluorescent imaging of amyloid aggregation, and further highlights the value of the naphthalimide

Open Access Article. Published on 05 July 2021. Downloaded 9/23/2021 12:29:10 PM. tetrazines that we have developed here. References

1 L. D. Lavis and R. T. Raines, ACS Chem. Biol., 2007, 3, Conclusions 142–155. 2 E. J. New, ACS Sens., 2016, 1, 328–333. Here we have presented the first naphthalimide tetrazines 3 J. L. Kolanowski, F. Liu and E. J. New, Chem. Soc. Rev., 2017, designed to undergo rapid click chemistry for biological label- 47, 195–208. ing. The suitability of naphthalimides for biological labelling 4 B. L. Oliveira, Z. Guo and G. J. L. Bernardes, Chem. Soc. Rev., events was assessed through the photophysical characteriza- 2017, 46, 4895–4950. tion of a series of these compounds, which found that Np6mTz 5 P. Shieh and C. R. Bertozzi, Org. Biomol. Chem., 2014, 12, had the best fluorogenic response in cuvette studies, and 9307–9320. underwent a fluorogenic reaction in cells with BCN within 6 C. S. McKay and M. G. Finn, Chem. Biol., 2014, 21, 20 min. The use of the naphthalimide scaﬀold permits the 1075–1101. synthetic variation of groups at the imide and naphthalene 7 D. M. Patterson, L. A. Nazarova and J. A. Prescher, positions, which led to the development of targeted naphtha- ACS Chem. Biol., 2014, 9, 592–605. limide tetrazines that fluorogenically labelled the lysosomes 8 M. L. Blackman, M. Royzen and J. M. Fox, J. Am. Chem. Soc., and mitochondria without genetic modification. The system 2008, 130, 13518–13519. was also used for no-wash imaging of insulin amyloid fibrils, 9 N. K. Devaraj, S. Hilderbrand, R. Upadhyay, R. Mazitschek which are an established model for understanding the progres- and R. Weissleder, Angew. Chem., Int. Ed., 2010, 49, sion of amyloidogenesis in amyloid-associated diseases. These 2869–2872. applications demonstrate the versatility of the naphthalimide 10 G. S. Jiao, L. H. Thoresen and K. Burgess, J. Am. Chem. Soc., scaﬀold and should pave the way for the design of new, 2003, 125, 14668–14669.

11 G. S. Jiao, L. H. Thoresen, T. G. Kim, W. C. Haaland, F. Gao, 31 A. Loredo, J. Tang, L. Wang, K. L. Wu, Z. Peng and H. Xiao, M. R. Topp, R. M. Hochstrasser, M. L. Metzker and Chem. Sci., 2020, 11, 4410–4415. K. Burgess, Chem. – Eur. J., 2006, 12, 7816–7826. 32 K. G. Leslie, D. Jacquemin, E. J. New and K. A. Jolliffe, Chem. 12 J. C. T. Carlson, L. G. Meimetis, S. A. Hilderbrand and – Eur. J., 2018, 24, 5569–5573. R. Weissleder, Angew. Chem., Int. Ed., 2013, 52, 6917–6920. 33 N. Trinh, K. A. Jolliffe and E. J. New, Angew. Chem., Int. Ed., 13 K. Lang, L. Davis, S. Wallace, M. Mahesh, D. J. Cox, 2020, 59, 20290–20301. M. L. Blackman, J. M. Fox and J. W. Chin, J. Am. Chem. 34 C. Le Droumaguet, C. Wang and Q. Wang, Chem. Soc. Rev., Soc., 2012, 134, 9–12. 2010, 39, 1233–1239. 14 B. Pinto-Pacheco, W. P. Carbery, S. Khan, D. B. Turner and 35 Q. Qiao, W. Liu, J. Chen, W. Zhou, W. Yin, L. Miao, J. Cui D. Buccella, Angew. Chem., Int. Ed., 2020, 59, 2–12. and Z. Xu, Dyes Pigm., 2017, 147, 327–333. 15 L. G. Meimetis, J. C. T. Carlson, R. J. Giedt, R. H. Kohler and 36 Z. Shao, C. Zhang, X. Zhu, Y. Wang, W. Xu, Y. Chen, R. Weissleder, Angew. Chem., Int. Ed., 2014, 53, 7531–7534. X. Wang, H. Zhu and Y. Liang, Chin. Chem. Lett., 2019, 30, 16 A. Wieczorek, P. Werther, J. Euchner and R. Wombacher, 2169–2172. Chem. Sci., 2017, 8, 1506–1510. 37 Y. Tian, X. Li and D. Yin, Chem. Commun., 2019, 55, 17 E. Kozma, G. E. Girona, G. Paci, E. A. Lemke and P. Kele, 12865–12868. Chem. Commun., 2017, 53, 6696–6699. 38 L. Fritea, K. Gorgy, A. Le Goff, P. Audebert, L. Galmiche, 18 G. Knorr, E. Kozma, J. M. Schaart, K. Ne´meth, G. To¨ro¨k and R. Sa˘ndulescu and S. Cosnier, J. Electroanal. Chem., 2016, P. Kele, Bioconjugate Chem., 2018, 29, 1312–1318. 781, 36–40. 19 G. Knorr, E. Kozma, A. Herner, E. A. Lemke and P. Kele, 39 A. J. Gross, R. Haddad, C. Travelet, E. Reynaud, P. Audebert, Chem. – Eur. J., 2016, 22, 8972–8979. R. Borsali and S. Cosnier, Langmuir, 2016, 32, 20 G. Beliu, A. J. Kurz, A. C. Kuhlemann, L. Behringer-Pliess, 11939–11945. M. Meub, N. Wolf, J. Seibel, Z. D. Shi, M. Schnermann, 40 O. Ourahmoun, T. Trigaud, B. Ratier, M. S. Belkaid,

Creative Commons Attribution-NonCommercial 3.0 Unported Licence. J. B. Grimm, L. D. Lavis, S. Doose and M. Sauer, L. Galmiche and P. Audebert, Synth. Met., 2017, 234, Commun. Biol., 2019, 2, 261. 106–110. 21 J. Tu, D. Svatunek, S. Parvez, A. C. Liu, B. J. Levandowski, 41 Q. Zhou, P. Audebert, G. Clavier, R. Meállet-Renault, H. J. Eckvahl, R. T. Peterson, K. N. Houk and R. M. Franzini, F. Miomandre, Z. Shaukat, T. T. Vu and J. Tang, J. Phys. Angew. Chem., Int. Ed., 2019, 58, 9043–9048. Chem. C, 2011, 115, 21899–21906. 22 A. Va´zquez, R. Dzijak, M. Dracˇ´ınsky´, R. Rampmaier, 42 L. D. Adair, N. Trinh, P. M. Ve´rite´, D. Jacquemin, S. J. Siegl and M. Vrabel, Angew. Chem., Int. Ed., 2017, 56, K. A. Jolliffe and E. J. New, Chem. – Eur. J., 2020, 26, 1334–1337. 10064–10071. 23 S. J. Siegl, J. Galeta, R. Dzijak, M. Dracˇ´ınskyánd M. Vrabel, 43 W. Mao, W. Shi, J. Li, D. Su, X. Wang, L. Zhang, L. Pan,

This article is licensed under a ChemPlusChem, 2019, 84, 493–497. X. Wu and H. Wu, Angew. Chem., Int. Ed., 2018, 58, 24 S. J. Siegl, A. Va´zquez, R. Dzijak, M. Dracˇ´ınsky´, J. Galeta, 1106–1109. R. Rampmaier, B. Klepeta´ˇrova´ and M. Vrabel, Chem. – Eur. 44 W. Chen, D. Wang, C. Dai, D. Hamelberg and B. Wang,

Open Access Article. Published on 05 July 2021. Downloaded 9/23/2021 12:29:10 PM. J., 2018, 24, 2426–2432. Chem. Commun., 2012, 48, 1736–1738. 25 S. J. Siegl, J. Galeta, R. Dzijak, A. Va´zquez, M. Del 45 D. Jacquemin, I. Duchemin and X. Blase, J. Chem. Theory Rı´o-Villanueva, M. Dracˇ´ınsky´and M. Vrabel, ChemBioChem, Comput., 2015, 11, 5340–5359. 2019, 20, 886–890. 46 A. Kaur, E. J. New and M. Sunde, ACS Sens., 2020, 5, 26 A. Egyed, A. Kormos, B. So¨veges, K. Ne´meth and P. Kele, 2268–2282. Bioorg. Med. Chem., 2020, 28, 115218. 47 E. Coelho-Cerqueira, A. S. Pinheiro and C. Follmer, Bioorg. 27 P. Werther, K. Yserentant, F. Braun, N. Kaltwasser, C. Popp, Med. Chem. Lett., 2014, 24, 3194–3198. M. Baalmann, D. P. Herten and R. Wombacher, Angew. 48 G. S. Waldo, B. M. Standish, J. Berendzen and Chem., Int. Ed., 2020, 59, 804–810. T. C. Terwilliger, Nat. Biotechnol., 1999, 17, 691–695. 28 F. Neubert, G. Beliu, U. Terpitz, C. Werner, C. Geis, M. Sauer 49 C. Wurth, N. K. Guimard and M. H. Hecht, J. Mol. Biol., and S. Doose, Angew. Chem., Int. Ed., 2018, 57, 16364–16369. 2002, 319, 1279–1290. 29 J. J. Gruskos, G. Zhang and D. Buccella, J. Am. Chem. Soc., 50 H. A. Shaban, C. A. Valades-Cruz, J. Savatier and S. Brasselet, 2016, 138, 14639–14649. Sci. Rep., 2017, 7, 1–10. 30 L. I. Selby, L. Aurelio, D. Yuen, B. Graham and 51 L. Nielsen, S. Frokjaer, J. Brange, V. N. Uversky and A. P. R. Johnston, ACS Sens., 2018, 3, 1182–1189. A. L. Fink, Biochemistry, 2001, 40, 8397–8409.